0

I am trying to configure a reverse proxy. I currently have

<Proxy balancer://myset>
    BalancerMember http://IP1:80 retry=4
    BalancerMember http://IP1:80
    Balancermember http://IP3:80
    ProxySet lbmethod=bytraffic
</Proxy>
ProxyPass "/balancer-manager" !
ProxyPass "/"  "balancer://myset/"
ProxyPassReverse "/"  "balancer://myset/"

By setting retry to 4, it lowers the time before the reverse proxy checks the health of a Member in the error state. However, if I run systemctl stop httpd (from the instance at IP1), when I refresh the balancer-manager page (of the reverse proxy), it takes a variable amount of time ( sometimes mere seconds) before it registers as Init Err. How do I configure this check to run less often? Is the check only running because I am refreshing the page? I am not using dynamic health checks, so is PING used and what's CPING, which is mentioned in the documentation? I have looked at the module page, which usually explains it exactly, but I have tried a few things, to no success. Also, if the addresses of the all BalancerMembers are pinged for every request, does that mean we have an extra 3 round trips for every request? Finally, if I decide to enable health checks (via hcheck module), does that disable the static checks? Thanks.

1
  • If I understand what you ask correctly, HTTPD tries the server when new requests come in, so the time before httpd notices if the backend server is down may vary, if you want to actively check the backend use mod_proxy_hcheck Nov 2 at 17:48

0

You must log in to answer this question.

Browse other questions tagged .